Voice Search Is The Future of SEO Services!

By: Adam Riggio, July 5, 2017

Afraid of Hearing Voices?

Real and Imagined Dangers of Voice Search for the Future of SEO Marketing

It’s not exactly hard today to find more and more content that claims to give you the one neat trick or a few simple solutions to the problems your business faces. That’s true in every field, SEO marketing included.

For the past couple of years, SEO marketing experts have been worrying that growth in voice search will cut into their results. But how many of these fears amount to the worries of paranoia? How much of that paranoia turns out to be right?

Convenience and Vulnerability

The main attraction of voice search and voice command in general is convenience. Many people want the ability to stick some earbuds into their smartphones and be able to direct it through a mic, without having to take it out of their pockets.

Yes, it’s convenient, but it can also be very dangerous. Imagine the panic that would ensue if you shouted “Siri! Order eight pizzas to my house!” or “Siri! Text mom to tell her I’m pregnant!” on a bus crowded with voice command users.

But voice users are vulnerable to more than humiliating pranks. Too often, voice commands are so easy for so many to use that you can forget the incredible differences between how humans and phones understand spoken words.

In our daily lives, we use the same ordinary language to refer to how people understand spoken words and how smartphone apps understand spoken words. That language literally is, “understanding spoken words.” Humans are social organisms whose entire personalities are developed through dynamic relationships whose main medium is language.

Siri and Google Assistant are machines made of silicon, metals, and electronic patterns. Their algorithms are very good at mimicking how we expect HAL 9000 or Mr. Data to sound. But they’re barely passable mimics of the actual behaviour and intelligence that those sci-fi role models can manage. Even the Enterprise’s computer is advanced far beyond any of our capabilities.

google search

Yet voice’s problems aren’t just matters of advancement. Thinking of voice command comprehension as a single process of learning and advancement mistakes how squishy carbon organisms learn for how metal and silicon machines learn.

One voice command security study puts paid to this category mistake. It’s possible to create audio clips that sound like a human command to a voice algorithm, but which sound like nonsensical demon screams or white noise to humans.

These audio files could be embedded into website design ads or youtube videos, so that when you play them on your smartphone, they activate commands and functions without any of your knowledge at all. That includes your confirmations of voice commands for sensitive orders like money transfers.

Even without accessing such secure activity, hidden voice commands can make your smartphone vulnerable to adware or malware.

Could There Be Limits to Voice?

However, there’s one important way in which many people are safe from this vulnerability. They simply don’t use them. The reason is that they have accents that most or all voice command systems can’t understand at all.

The most common way of discussing this issue is to mine laughs from it. Just look at a sketch like this, where two people with heavy Scottish accents are stuck in an elevator because the voice recognition software can’t interpret their words properly. “Oh, those Scotsmen! They’re so funny when they’re enraged!”

Now, the sketch is nearly a decade old – that video was uploaded to YouTube in 2010. But voice command problems continue. Just a few months ago at Wired, Sonia Paul wrote about how many accents still can’t interact with voice commands at all.

It’s almost as if most voice recognition software was designed only to recognize anglophone-from-birth accents from northern California. I wonder why that is? he asked sarcastically.

But it’s a serious practical problem for voice search. For one thing, the United States and Canada are diverse countries. People from all over the world have come to North America both as immigrants and as visiting businesspeople.

Think about the limitations Californian voice command developers would face promoting their products around the world if voice recognition can’t handle anyone other than an English-speaker-for-life?

If Paul’s Filipino mother and Indian father can’t use them after speaking English fluently for decades, what use would a Californian voice system be for other languages entirely?

Can Voice Catch Up With Changing Voices?

Voice search’s accent problem could even worsen as American English accents diversify. Over the last few generations, the accents of white Americans in the Great Lakes region have shifted how they pronounce key vowels. The pronunciations of the Great Lakes vowel shift haven’t changed since 800AD.

Will the still-problematic voice command systems of Silicon Valley’s hype machine be able to cope with an America where millions of people have accents almost as diverse as Europe or India?

Ultimately, however, the problem of non-white, non-Californian accents is something that engineers can power through. It’s a matter of building increasingly complex voice comprehension algorithms to deal with the human voice’s full range of diversity across languages and accents.

If I can describe the difficulty of this task in technical terms, that’s going to take a whole hell of a lot of work. But it will mean voice command applications will be able to act on a wider variety of human sounds – and human-like sounds.

So the security vulnerabilities of highjacking voice commands with sounds that humans may not understand, but that voice command software can, will grow as the software can comprehend more sounds.

The continuing development of voice command software must and will go hand in hand with the development of security protocols. As voice can take action based on more sounds and more diverse sounds, there will be more possible hacks through carefully organized white noise.

That need to maintain security hand in hand with flexibility is the most important problem holding the development of voice command software back from total ubiquity in societies across the globe.

If All This Is Solved, Can Voice Still Bleed SEO?

Despite these far-reaching and complicated problems, there remains a more fundamental issue of voice that can still strike fear into an SEO marketer. To understand that issue, we’ll have to take a very profound look at SEO and voice, to see how they work as media.

So what is the medium of SEO? The light from a screen. Maybe I’m being a little too literal when I talk about SEO as a medium. But that’s what makes voice seem such a threat.

SEO is all about moving your company’s website presence up in the rankings of relevant search results. When you type a query on a screen, a variety of results appear for you to choose from.

Each question has tens or hundreds of possible answers. SEO makes your answers to consumers’ questions more prominent in that list. But the ultimate choice remains with the consumer, with the person who asked the question in the first place.

Voice can easily take that power away from the consumer because when you ask a question, you get one single answer.

“Siri, what’s a good Greek restaurant near me?” And Siri will just tell you the answer.

SEO rankings are useless when people aren’t presented with many alternatives to search through.

What Is Voice For?

Voice search is based on an appeal to convenience, but you can just as easily understand our need for convenience as humanity’s fundamental laziness.

When you don’t want to do any work and you just want to be told the answer, voice search works fine. Laziness is a powerful motivator, but it may not be powerful enough to replace typed search queries entirely, or even by all that much.

Here’s why. The three most popular smartphone voice commands are 1) “Navigate home,” 2) “Call mom,” and 3) “Call dad.” What can this tell us about how people use voice day to day?

It tells us that a voice search isn’t really a search for people. It’s a demand for action. The decision has already been made.

Here’s another thing about most voice searches. They’re simple and casual. Ordinary requests and questions to which you already know the answer. When I’m asking about that Greek restaurant, maybe I have a few particulars in mind, or maybe I’m going to look for the cheapest one. Maybe I’ll just be led to the only one in the neighbourhood.

But if my situation is any more complex than that, I won’t want an easy answer. An easy answer would only confuse me. So when Siri tells me there are a lot of restaurants in the area and begins listing addresses, Yelp scores, articles about them on local blogs, I won’t want to hear all that in a monologue.

I’ll ask a question that’s music to an SEO marketer’s ears. “For God’s sake, Siri, just show me the search results! I’ll read them myself!”